Standardizing Safety Assessment and Reporting for Neonatal Clinical Trials

Despite numerous pediatric legislative initiatives worldwide, neonates are largely treated with medicines that have not been specifically approved or indicated for this population.^1–5 Some of the reasons cited for why so few products have been developed for neonates include gaps in understanding complex pathophysiology, a small market with unique conditions, challenges in the design and execution of clinical trials, and difficulties with the assessment of safety and efficacy. Evaluation of drug safety in neonates is frequently complicated by exposure to multiple drugs, high rates of comorbidities, and complex assessments of short- and long-term outcomes. All of these issues may discourage investigators and sponsors from studying therapies in neonates, who remain a highly underserved population. It has been over 25 years since a novel medicine such as surfactant has been specifically developed for premature neonates to improve their survival and outcome.

To facilitate drug development efforts in neonates, the Critical Path Institute and the Food and Drug Administration (FDA) established the International Neonatal Consortium in 2015. This organization has convened key stakeholders from the academic community, industry, global regulatory agencies, parent advocates, nursing groups, and funding agencies with a focus on regulatory science. The International Neonatal Consortium aims to develop approaches to standardize and harmonize efforts to define and measure safety and efficacy of medications used in neonates.⁶ Clarifying what is considered a safety endpoint and an efficacy endpoint is important because neonatal morbidities and mortality are often integral to both categories. Complete assessment and reporting of adverse events (AEs) are essential to understanding the potential risks of any intervention. A well-informed benefit-risk assessment is essential to support progress from early phase studies to larger pivotal trials, with the ultimate goal of approval of safe and effective therapies for neonates in multiple jurisdictions.

The current systems for ascertaining (the research staff often learns about a potential AE from the clinical staff caring for the neonate) and adjudicating AEs are heterogeneous and burdensome for sponsors and investigators. Consistent approaches are needed as academic trialists, contract research organizations (CROs), and industry sponsors may have their own individual strategies for monitoring safety. CRO and/or industry staff tasked with monitoring the study may have limited knowledge of neonatal pathophysiology, pharmacology, and toxicology which makes standardizing the definitions and approach even more important. It is often difficult for regulators, Institutional Review Boards (IRBs), Ethics Committees, and Data Safety Monitoring Boards (DSMBs) to evaluate whether AEs are likely drug-related, the result of drug-drug interactions, or related to underlying disease processes (often extreme prematurity). Background rates of morbidities and mortality, which may inform these assessments, differ greatly across institutions and regions of the world. These complexities have resulted in extraneous documentation and excessive burden on research staff, which can deter investigators from conducting neonatal clinical trials. The Tufts Center for the Study of Drug Development has reported that industry protocols have become significantly more complex, often placing unreasonable demands and burdens on all staff at recruiting sites.⁷

Innovative approaches have been used to address safety reporting in other populations and are urgently needed in neonates.^8,9 This commentary will first review definitions and existing regulatory requirements for safety reporting. Next, approaches to standardize charting and coding to optimize AE reporting will be provided. How safety monitoring review boards and global regulatory agencies review and address AEs will also be explored. Finally, potential approaches for standardizing and streamlining safety reporting for neonatal clinical trials will be presented.

Regulatory Definitions, Requirements, and Considerations

The European Commission clinical trial safety guidance was updated in 2011.¹⁰ FDA Guidance published in December of 2012 outlined requirements for investigational new drug safety reporting for human drug and biological products.¹¹ A follow-on Draft Guidance was published in 2015, providing a framework for safety assessments.¹² The Pharmaceuticals and Medical Devices Agency in Japan and Health Canada follow similar guidance to the US and European Union with the definitions listed below used by most international regulatory bodies (except where noted). The following definitions and requirements are general and do not specifically relate to neonatal clinical trials.

An AE is defined as any untoward medical occurrence in a patient or clinical trial participant, administered or exposed to a medicinal product, and which does not necessarily have a causal relationship with the product. A suspected adverse reaction (SAR) means any AE for which there is a reasonable possibility that the investigational drug caused the AE. SAR implies a lesser degree of certainty than an adverse reaction, which is any AE caused by a drug. “Reasonable possibility” means there is evidence to suggest a causal relationship between the drug and the AE. A life-threatening AE or life-threatening SAR has occurred if in the view of the sponsor or investigator the event placed the patient at immediate risk of death. An AE is considered a serious AE or serious SAR if in the view of the investigator or sponsor it results in any of the following outcomes: death, life-threatening, inpatient hospitalization or prolongation of hospitalization, persistent or significant incapacity or disruption of normal life functions, or a congenital anomaly. Important medical events that may also be considered serious are those that jeopardize the patient and may require medical or surgical intervention to prevent one of the above outcomes. An unexpected AE or unexpected SAR is one that is not listed in the investigator brochure or reference safety information or is more specific or severe than what was previously observed. A serious and unexpected suspected adverse reaction (SUSAR) is an event that is assessed to be causally related to the investigational drug, is not listed in the available safety information, and is deemed serious. In the European Union, these are reported to the national competent authority of the Member State concerned. In the US, SUSARs are reported to the FDA if the sponsor determines there is reasonable possibility of causality.

The fundamental principles of safety reporting in US regulations/guidance, European Commission guidance, Canadian Regulations, and Japan’s Ordinance for Enforcement of the Pharmaceutical Affairs Act are that sponsors and investigators should perform an ongoing safety evaluation throughout each clinical trial. This would include reviewing and assessing SAEs, submitting unexpected serious adverse drug reactions, and periodically reviewing safety data that has accumulated across all studies and external sources. Sponsor responsibilities include surveillance of all relevant sources of information and reporting to authorities and investigators. This surveillance includes any serious and unexpected AE such as a single occurrence of an uncommon event that is strongly associated with drug exposure, one or more occurrences of an event that is unusual in the exposed population but not commonly associated with drug exposure, or analysis of specific AEs observed in the trial that appear more common in the study drug-exposed group than in the control group. Findings from epidemiologic studies, meta-analyses, or other clinical studies that suggest a significant risk in humans exposed to a drug, findings from animal or in vitro testing that suggest a significant human risk such as mutagenicity, teratogenicity, carcinogenicity, or significant organ toxicity at or near human exposure levels, or increased rate of occurrence of serious adverse drug reactions are also important. The timelines for reporting new information to regulatory agencies and to investigators vary between jurisdictions. A consistent timeline exists for reporting SUS-ARs to regulatory agencies; within 7 days if the SUSAR is fatal or life-threatening or 15 days for events meeting other criteria for a serious event.

Unanticipated problems should be reported to the IRB in a timely fashion.¹¹ Individual investigational new drug safety reports should be submitted to the review section of the regulatory body. The reporting requirements apply to investigator-initiated as well as sponsor-initiated studies. Regulators recognize that individual investigators do not have access to the complete data maintained by a commercial sponsor, but investigators are required to evaluate all safety information that is available to them to determine whether any events qualify for reporting. In addition, sponsors, investigators, and DSMBs should examine the literature and actively track any new safety information about the product and/or population under investigation.^11–13

Clinical safety endpoints, both short- and long-term, may not be known at the initiation of a study if the trial is a first in human study in neonates. Any safety data from animal studies, adult human studies, and from drugs in the same/ similar class with a similar mechanism of action should be leveraged to design the data safety monitoring plan for the clinical trial.¹⁴ Blinded and randomized controlled trials (whenever ethical and feasible) are fundamental to assess drug safety given the complexities inherent in the population. It also is important to understand the use of concomitant medications (although not necessarily documenting each dose, dosing interval, and route of administration) and any potential drug-drug interactions. Understanding the mechanism of action of the investigational drug is particularly important as the ontogeny of organ, tissue, and cellular/ receptor systems may impact the safety profile of the therapeutic product being evaluated. This may also help investigators anticipate what type of AEs could potentially occur.

Once a study is underway, it is important to report any AEs that appear to be related to the therapeutic product, especially if there is a dose-response relationship or evidence of causality (eg, from a dechallenge-rechallenge trial). Serious AEs (including deaths) all need to be reported and require detailed analysis of comprehensive narratives, including autopsy findings if available. Although the relationship between the therapeutic product and the AE may become more difficult to assess over time, any serious AEs or deaths during the initial hospitalization should be examined for potential causality. If the product or its class may be associated with long-term safety concerns, careful follow-up after discharge from the neonatal intensive care unit should be performed.

Safety Reporting from the Industry and CRO Perspective

Neonatal drug development represents a significant challenge for pharmaceutical companies and CROs.¹⁵ This is primarily due to the (1) severity of neonatal illness; (2) frequent comorbidities; (3) exposure to multiple drugs; (4) lack of established normal reference ranges for laboratory values; (5) limited numbers of new treatments developed specifically for this population; (6) difficulty with selecting the best dose and dosing interval; (7) limited number of qualified sites; and (8) limited number of experienced investigators. Although all of these factors contribute to difficulties with safety reporting, it is essential to change this paradigm and promote a multidisciplinary approach among all key stakeholders from the earliest stages of drug development.

Determining causality of AEs in the neonatal population is more difficult due to inconsistent terminology and incomplete descriptions at the site level. Although sponsors use Medical Dictionary for Regulatory Activities (Med-DRA) terminology for AE reporting, multiple terms and definitions may be used for similar events (eg, neonatal sepsis or suspected sepsis and changing definitions and classification of bronchopulmonary dysplasia), which contribute to inconsistent reporting. Whenever possible, a diagnosis rather than a list of signs should be recorded. If a diagnosis has not been made or a recognized syndrome identified, then each key sign should be listed individually. Many sponsors have now created their own case report forms to ensure that the AEs are captured and classified accurately. Finally, there can be clusters of reportable AEs such as pneumothorax, bradycardia, and hypoxemia that may all be related to a single AE. Because there is no consistent process for determining the principal AE, all the individual signs or events are reported independently and counted as separate AEs.

Sponsors are required to report all SARs. One approach is an aggregate analysis of specific events observed in a clinical trial (eg, known consequences of the underlying disease or condition under investigation or other events that commonly occur in the study population independent of drug therapy) that indicates the events occur more frequently in the drug treatment group than in a concurrent or historical control group. Because there are significant background rates of comorbidities and a lack of standard definitions of AEs, these assessments may be particularly challenging.

A sponsor or investigator’s ability to assess long-term safety is also difficult. Compared with treatment effects in adults, neonatal trials may involve drugs that have complex effects on multiple organ systems at a stage where drugs rarely have precisely targeted effects. Some effects cannot be detected or measured (by accepted or validated modalities) for years and are subject to multiple environmental influences that may confound the results of clinical trials. Feasibility of long-term safety follow-up in neonates has been problematic because of high dropout rates, children lost to follow-up, and increased costs associated with tracking these children.¹⁶

Standardizing Charting and Coding to Ensure Safety and Quality

According to the Center for Disease Control, adoption of electronic medical records (EMR) into healthcare is currently >80% in the US, but is more variable in other parts of the world.¹⁷ The EMR can be an important source of AE information. However, because clinicians enter information in the medical record in different manners, it is often difficult to extract narrative information in a coded fashion. There are significant problems using the EMR for clinical data including (1) variability in the quality and accuracy of data entry; (2) lack of retrieval systems for input of free text (not entered in a standardized fashion); (3) lack of common definitions for many disease processes; (4) lack of common data elements; (5) multiple different health care personnel, with varying levels of expertise who are entering data; and (6) disease processes and diagnoses that change over time. In the current paradigm, clinical research coordinators review the EMR and manually input data into the trial reporting database.

AE Causality Assessment and Severity Grading

Pharmacovigilance is based primarily on prevention, detection, and assessment of AEs which must be adapted for neonates. There are regulatory requirements to assess seriousness and causality during drug development. Population-tailored tools to evaluate seriousness/severity and causality are required to further facilitate neonatal drug development.^18,19

Following detection, disentangling “true” acute drug reactions from confounding events (eg, maturational changes in laboratory values, organ dysfunction, comorbidities) remains a challenge.¹⁵ Because of these confounding variables, commonly applied adverse drug event algorithm scoring systems like the Naranjo algorithm or the World Health Organization’s Uppsala Monitoring Center causality tool do not sufficiently and reliably document causality in neonates.^20,21 Although none of these scores have been validated in prospective clinical trials, a population specific tool to assess causality in neonates has been reported.²⁰

Du et al developed and subsequently validated an algorithm using a stepwise and systematic approach to better define acute drug reactions in the neonatal intensive care unit setting.²⁰ First, experienced pediatric pharmacologists used a 24-item questionnaire to rate 100 suspected AEs as definite, probable, possible, or unlikely with subsequent discussion to reach agreement. Based on this, a new tool was developed based on 13 questions (yes/no/not applicable) that could be quantified (≤2, 3–6, 7–13, ≥14) and categorized (definite, probable, possible, unlikely, not related).²⁰ This tool was subsequently validated using the new Naranjo tool and was demonstrated to be more reliable. Although further prospective testing is needed, this type of approach may be useful for surveillance and assessment in neonates.

Severity grading is commonly subdivided into mild, moderate, severe, life threatening, or death (grade 1–5). AE severity scales offer guidance on how to assign a severity grade to an AE. They are consensus documents, with the intention to reduce or limit inter-rater variability. Various scales exist and are commonly related to specific topic interests such as infectious diseases, vaccine trials, or oncology^8,9,18,19; but none of these existing grading systems are compatible with assessments in neonates and a more specific systematic approach is urgently needed.

Human Subjects Protections and the Roles of Safety Monitoring Review Boards and Global Regulatory Agencies

In the US, clinical research involving children must follow Health and Human Services “Additional Protections for Children Involved as Subjects in Research” (45 CFR 46 Subpart D),²² which are parallel to FDA regulations 21 CFR 50, that are invoked for research involving FDA-regulated products. Both regulations indicate that each IRB “must review clinical investigations involving children covered by subpart D and approve only those clinical investigations that satisfy the criteria of all applicable sections.” (1) Approval can only occur when the clinical investigation is no greater than minimal risk; (2) Investigations involving greater than minimal risk must provide the prospect of direct benefit to the subject. The risk must also be justified by the anticipated benefits with the relation of the anticipated benefit to the risk at least as favorable to subjects as that provided by available alternative approaches; and (3) Investigations with greater than minimal risk and no prospect of direct benefit are permitted if the intervention or procedure is likely to yield generalizable knowledge about the neonate’s disorder or condition which is of vital importance for the understanding or amelioration of the disorder or condition.

An essential element for each of these sections is permission of the parents or guardians because informed consent for neonatal studies involves additional challenges. Parents may be asked to participate in trials under difficult circumstances such as premature birth or other complications of labor and delivery. They may be asked to make decisions quickly, as the timing of the intervention may be critical. If the neonate is transported to a hospital with a higher level of care, they may not be able to remain with their neonate. Parents may be frightened and isolated and dependent on safeguards for neonates in clinical investigations to protect their family.

In addition, FDA issued a guidance document in 2012 on IRB Continuing Review after Clinical Investigation Approval.¹³ One important protection for vulnerable populations is the use of DSMBs. Clinical investigators and IRBs are required to analyze the risk and benefit of each clinical trial. For trials of therapeutic products for neonates, data may not be available in adults or older children to allow a robust analysis of risk and benefit. It is important that complete, relevant safety data be collected throughout the clinical trial and DSMBs need to carefully and periodically review these data in the context of the trial. FDA has provided guidance on the establishment and operation of these Clinical Trial Data Monitoring Committees.²³ A fundamental reason to establish a DSMB is to enhance the safety of trial participants when safety concerns may be unusually high, especially if the study is being performed in a population at elevated risk of death or other serious outcomes. Based on these criteria, all blinded therapeutic intervention trials in neonates should employ a DSMB. Nonblinded intervention studies may be able to rely on an independent safety committee for ongoing reviews. Perrem et al examined multiple neonatal randomized controlled trials and found that 61% of trials used a DSMB.²⁴ However, only 51% of the reports mentioned an interim analysis and 21% reported stopping rules. This indicates that independent safety monitoring should be an important aspect of randomized controlled trials in neonates and there should be clear documentation of safety-related issues.

Once a therapeutic product has been approved and labeled for use in neonates, it is important to continue to track short-and long-term safety. Many times, it is not possible to identify rare SAEs in premarket studies because of the limited numbers of neonates enrolled in the trials. Once a product is marketed and use increases, new short and long-term AEs may be identified. Appropriately designed neonatal registries could help capture this information so that additional postmarketing safety data could be added to product labeling.

Clinical trials are increasingly global in scope. As such, it is important that AEs in neonatal clinical trials are reported globally. On a monthly basis, there is a pediatric cluster discussion among international regulatory agencies including FDA, European Medicines Agency, Health Canada, Australia’s Therapeutic Goods Administration, and Japan’s Pharmaceuticals and Medical Devices Agency. Safety issues may be identified and discussed by all key stakeholders as needed in this forum.

Proposed Solutions

When studying any new or existing drug, safety assessments require comprehensive monitoring, collecting, and reporting of data. Standardized protocol designs and data collection tools are urgently needed to obtain data of sufficient quality for regulatory submissions.

Standardization of EMR Data Entry and Preparation of Structured Datasets

All data to be collected and entered should be listed, structured, and coded in advance. Clinical staff will require training on data standards so that they may select the correct and specific clinical findings from dropdown menus. These actions may facilitate the use of more precise terms, which will serve to enhance data quality. A focus on automated data extraction from the EMR using structured and coded terminology (eg, disease name) is needed with incorporation into ongoing quality improvement programs.²⁵ As EMR algorithms evolve, they could be a reliable data source for AE reporting, quality improvement initiatives, and risk-based monitoring. The ultimate goal would be to build upon EMR technology to provide more consistent and efficient trial safety monitoring.

Clarification of Neonatal Protocols

The study protocol, designed in collaboration with multiple key stakeholders (including parents), should have a comprehensive safety section clearly delineating which items need to be reported and in what timeframes. It is important to decide a priori whether some or all concomitant medications (indication, dose, dosing interval, duration of treatment) need to be recorded.

Specific Neonatal AE Severity Scales

A standardized Neonatal AE Severity Scale should be developed to provide consistent terminology to report AEs. All AE terms in the scale should be linked to the lowest level terms in MedDRA, codes of the NCI Thesaurus Pediatric Adverse Event Terminology Subset, and other international dictionaries while meeting harmonization and submission standards needed for regulatory submissions. Investigators should report an AE with as much detail as possible, so that the sponsor and/or CRO can link it to a consistent Med-DRA term. Reports that are nonspecific or unclear will result in inconsistent MedDRA coding and additional queries to the study staff that will add to the burden of reporting for a clinical trial.

A stepwise, multidisciplinary consensus-driven Delphi approach to development of a neonatal AE severity scale is needed with plans for subsequent validation. Existing grading systems based on “activities of daily life and selfcare” as severity markers are not applicable to neonates. Consensus on maturational activities of daily life and related generic severity grading criteria will be a first step, and neonatal AE specific scales should include neurological, cardiovascular, respiratory, gastrointestinal, and infectious events. The scope and type of events should also be part of the Delphi effort. Once developed, these scales should be validated, preferably embedded in ongoing clinical trials.

Medical Monitoring

Neonatal trials can involve many global sites, requiring medical monitors to review source documents and assist with AE reporting. It is critically important to have investigators, support staff (eg, study nurses, study coordinators), and study monitoring staff (representing the sponsor) with sufficient expertise in neonatal clinical trials to review all potential AEs carefully and efficiently. Because enrollment in these trials may take 3–5 years or longer (especially with longer term follow-up needed for neonates), turnover of study staff can be particularly problematic. Over the past 2 decades, the pharmaceutical industry and CROs have gained significant experience in managing pediatric trials, yet medical monitors and other personnel knowledgeable in neonatal medicine are rare, which can add burdens for the site. It is essential to focus on the long-term needs of industry and CROs and increase the number of staff specifically trained in pediatrics and neonatal medicine to meet the demand.

Collaborative Approaches to Address Challenges in AE Reporting

Sponsors, investigators, CROs, and regulatory/oversight boards all struggle with AE reporting in neonatal clinical trials. Through collaborative efforts, sponsors are establishing large databases containing study participants from placebo control and standard of care arms from completed adult clinical trials. This is designed to enhance safety signal detection by having a better understanding of background incidence rates of important morbidities by region of the world.²⁶ These efforts are currently limited by relevant International Classification of Diseases and Related Health Problems, 9th and 10th revision coding guidelines. The neonatology community may collaborate or learn from these ongoing efforts and propose a similar solution. Additional collaborative efforts should be directed to develop Core Neonatal Datasets outlining standardized data elements and definitions to be included in all neonatal clinical trials.

Summary and Conclusions

Assessing the safety of new and existing drugs in neonates is critically important and extremely difficult. All stakeholders are excessively burdened by the current system, which is less than optimal for neonates, who may be exposed to medicines that have not been adequately evaluated and for whom there are unmet therapeutic needs. Although safety evaluations in neonates involve the application of standard pharmacovigilance principles and practice, additional challenges exist in this population because of a high background rate of serious complications, which can interfere with establishing a direct relationship to the therapeutic agent under study. There is a need to standardize and harmonize reporting requirements, reporting procedures, and data collection as well as assessments, interpretation, and review of AEs. The direct financial costs of clinical trial safety assessments should be recognized, and sites should be appropriately reimbursed. In addition, the burden on sponsors and investigators to monitor and collect data is expensive and time consuming. These issues can serve to discourage much needed research for this underserved patient population.

Although it may be challenging to prespecify safety endpoints for neonatal clinical trials, it is important to leverage all available safety information to design a robust safety evaluation that includes both short-term and long-term assessments. The workforce needs a better balance of neonatology and pharmacovigilance expertise. Sponsors and CROs require expertise and competence in neonatal medicine to successfully design, operationalize, and execute neonatal clinical trials. In addition, the academic neonatal community (both clinical and research) requires a broader understanding of the regulated drug development process. Collaborative efforts of all key stakeholders, including clinicians, researchers, industry, regulators, and patient advocates, are essential for developing protocols and safety assessment tools, building data standards and reference ranges, and understanding background AE rates to streamline and standardize safety reporting.

Strategic planning is needed to gather expertise and resources to support activities such as building consensus, collecting and analyzing data, and piloting and validating assessments. New technologies and harmonized electronic data capture need concerted development. Companies are currently developing facilitated risk monitoring tools and practices that will need to be applied to the neonatal population across the globe. We call on all stakeholders to contribute to the development and implementation of a comprehensive plan to standardize, harmonize, and greatly improve the evaluation of safety of medicines used in neonates.

Glossary

AE

Adverse event

CRO

Contract research organization

DSMB

Data Safety Monitoring Board

EMR

Electronic medical record

FDA

Food and Drug Administration

IRB

Institutional Review Board

MedDRA

Medical Dictionary for Regulatory Activities

SUSAR

Suspected adverse reaction